Surgical mapping of middle ear cholesteatoma with fusion of computed tomography and diffusion-weighted magnetic resonance images: Diagnostic performance and interobserver agreement.

OBJECTIVE

To assess the diagnostic performance in detecting primary cholesteatoma at various anatomical subsites using Computed Tomography (CT), Diffusion-weighted Magnetic Resonance Imaging (DWMRI) and Fusion of CT and DWMRI (Fusion CT-MRI) images.

STUDY

DESIGN

A retrospective study of 22 children identified from a prospective database of surgically treated cholesteatoma cases over a five year period. All cases underwent pre-operative CT, non-echo planar DWMRI and Fusion CT-DWMRI, and with clearly documented surgical findings. For each imaging modality, two radiologists scored for the presence or absence of cholesteatoma with confidence levels at different anatomical subsites. The radiologists were blinded to the surgical findings to which their findings were compared.

SETTING

Large Teaching Hospital in London.

PATIENTS: 22 children with cholesteatoma confirmed surgically.

INTERVENTION: CT, DWMRI imaging and fusion CT-MRI.

MAIN OUTCOME MEASURE: Diagnostic performance of subsite localisation of cholesteatoma by CT, DWMRI and fusion CT-MRI imaging with intra-operative findings.

RESULTS

Twenty-two patients were included (12 women and 10 men). The median age of patients was 11 years. When considering all subsites combined, the result for all imaging methods suggested 'good' agreement between both observers. When all subsites were examined together, all methods had relatively high sensitivity values (87% for CT vs 84% for DWMRI vs 85% for fusion CT-DWMRI). Specificity was highest with fusion CT-DWMRI (46% for CT vs 76% for DWMRI vs 97% for fusion CT-DWMRI), as was accuracy (66% for CT vs 80% for DWMRI vs 91% for fusion).

CONCLUSIONS

Our study has demonstrated that fusion CT-DWMRI is superior to DWMRI or CT separately in localizing cholesteatoma at various middle ear cleft subsites and bony relations, making it a valuable tool for surgical planning.